On December 15th, NVIDIA (NVDA.O) made headlines by unveiling a range of new open-source artificial intelligence models on Monday. The tech giant asserts that these cutting-edge models surpass their predecessors in terms of speed, affordability, and intelligence. NVIDIA is well-known for its chip supply prowess, with companies such as OpenAI leveraging its chips to train closed-source models and reap profits. Yet, NVIDIA is not resting on its laurels; it is also vigorously developing its own models that span diverse domains like physical simulation and autonomous vehicles. These models are being offered in an open-source manner, enabling researchers and businesses to harness their potential. For example, Palantir Technologies has already seamlessly integrated NVIDIA's models into its offerings. On Monday, NVIDIA introduced the third iteration of its 'Nemotron' large language model, tailored for tasks like writing and programming. Notably, the smallest-scale Nemotron 3 Nano was released concurrently, with plans to unveil two larger-scale versions in the first half of 2026. Furthermore, reports suggest that Meta Platforms (META.O) is contemplating a shift to closed-source models, a move that could potentially elevate NVIDIA to a prominent position among major suppliers in the U.S. open-source model landscape.